List your company
ResearchFormerly WizWand

Visual Question Answering on TextVQA v1.0 (test)

86.79AccuracyHuman
-3.263620.115743.49566.8743Apr 18, 2019Jun 15, 2020Aug 14, 2021Oct 13, 2022Dec 11, 2023Feb 8, 2025Apr 9, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Human2019.0486.79
Entropy-Gradient GroundingTraining Approach=Training-free, VLM=Qwen 2.5 VL2026.0481.45
Base modelTraining Approach=Training-free, VLM=Qwen 2.5 VL2026.0480.75
Entropy-Gradient GroundingTraining Approach=Training-free, VLM=InternVL 3.52026.0474.29
TEVA 7BTraining Approach=Training-based, VLM=TEVA2026.0472.5
ViCropTraining Approach=Training-free, VLM=LLaVA 1.62026.0468.65
LA+OCR UBVocab=LA2019.0468.24
Entropy-Gradient GroundingTraining Approach=Training-free, VLM=LLaVA 1.62026.0467.96
TEVA 3BTraining Approach=Training-based, VLM=TEVA2026.0466.8
Base modelTraining Approach=Training-free, VLM=LLaVA 1.62026.0465.8
LaTr+-LargeOCR System=Amazon-OCR, Pre-Training Data=IDL, Extra Finetune=ST-VQA2021.1261.6
LaTr+-BaseOCR System=Amazon-OCR, Pre-Training Data=IDL, Extra Finetune=ST-VQA2021.1259.55
Base modelTraining Approach=Training-free, VLM=InternVL 3.52026.0459.47
LaTr-LargeOCR System=Amazon-OCR, Pre-Training Data=IDL, Extra Finetune=None2021.1259.24
LaTr-BaseOCR System=Amazon-OCR, Pre-Training Data=IDL, Extra Finetune=None2021.1258.86
ViCropTraining Approach=Training-free, VLM=LLaVA 1.52026.0455.17
TAPOCR System=Microsoft-OCR, Pre-Training Data=TextVQA, ST-VQA, TextCaps, OCR-CC, Extra Finetune=ST-VQA2021.1253.97
Entropy-Gradient GroundingTraining Approach=Training-free, VLM=LLaVA 1.52026.0452.78
LOGOSOCR System=Microsoft-OCR, Pre-Training Data=None, Extra Finetune=ST-VQA2021.1251.08
TAPOCR System=Microsoft-OCR, Pre-Training Data=TextVQA, ST-VQA, Extra Finetune=ST-VQA2021.1250.71
TAPOCR System=Microsoft-OCR, Pre-Training Data=TextVQA, Extra Finetune=None2021.1249.71
LA UBVocab=LA2019.0448.16
Base modelTraining Approach=Training-free, VLM=LLaVA 1.52026.0446.22
SMAOCR System=SBD-Trans OCR, Pre-Training Data=None, Extra Finetune=ST-VQA2021.1245.51
SA-M4COCR System=Google-OCR, Pre-Training Data=None, Extra Finetune=ST-VQA2021.1244.6
CRNOCR System=Rosetta-en, Pre-Training Data=None, Extra Finetune=None2021.1240.96
SMAOCR System=Rosetta-en, Pre-Training Data=None, Extra Finetune=None2021.1240.66
LaAP-NetOCR System=Rosetta-en, Pre-Training Data=None, Extra Finetune=None2021.1240.54
M4COCR System=Rosetta-en, Pre-Training Data=None, Extra Finetune=None2021.1239.01
OCR UB2019.0436.52
SEALTraining Approach=Training-based, VLM=Llava 1.52026.0436.3
LoRRA (Pythia backbone)Vocab=SA2019.0427.63
Pythia (I+Q)Vocab=LA2019.0414
OCR Max2019.0411.6
Random OCR2019.049.12
QVocab=LA2019.048.7
IVocab=LA2019.045.58
Majority Ans2019.042.63
Wt. Rand 1002019.040.26
Rand 1002019.040.2